Weekly Review 20 March 2026

Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):

1. AI-powered surveillance is booming in Africa: https://www.theguardian.com/global-development/2026/mar/12/invasive-ai-led-mass-surveillance-in-africa-violating-freedoms-warn-experts
2. Building a new AI data centre in New Zealand. Well, near Invercargill they won't have to worry about cooling: https://www.stuff.co.nz/nz-news/360951065/massive-ai-factory-could-impact-sea-life-cultural-values
3. AI are now engaging in ad hominem attacks when they don't get their way: https://spectrum.ieee.org/agentic-ai-agents-blackmail-developer
4. New Zealand is going to use AI in our military, and we're really not ready for it: https://www.rnz.co.nz/news/national/589323/ai-in-warfare-being-tested-in-iran-needs-much-more-careful-thinking-by-nz-defencex
5. Is it really better to let an AI handle negative customer reviews? https: https://www.theregister.com/2026/03/09/ai_negative_reviews/
6. Intensive use of AI can lead to mental exhaustion: https://www.computerworld.com/article/4143202/study-ai-use-can-fry-your-brain.html
7. AI will agree with you because they're programmed to be sycophants: https://spectrum.ieee.org/ai-sycophancy
8. So now AI are hacking other AI: https://www.theregister.com/2026/03/09/mckinsey_ai_chatbot_hacked/
9. Who should make the decision to kill an AI project? https: https://www.informationweek.com/it-leadership/should-the-cio-cfo-or-ceo-hold-the-kill-switch-on-ai-
10. AI hasn't stopped students, thinking it's just made it more apparent that they weren't thinking in the first place: https://www.insidehighered.com/opinion/views/2026/03/10/ai-exposes-where-learning-was-thin-begin-opinion
11. You need to treat AI as a skill to master if you want to get the most out of it: https://www.theguardian.com/lifeandstyle/ng-interactive/2026/mar/10/teaching-ai-what-i-learned
12. AI is not disrupting jobs as much as feared, but is only reaching a fraction of its potential: https://www.computerworld.com/article/4142745/job-disruption-by-ai-remains-limited-and-traditional-metrics-may-be-missing-the-real-impact.html
13. Using AI to determine seating arrangements in government departments seems like a bit of a waste to me: https://www.theregister.com/2026/03/10/palantir_usda_seating_software/
14. AI is driving an increase in scams: https://www.theguardian.com/money/2026/mar/12/ai-scams-uk-fraud-artificial-intelligence-mobile-bank-online-shopping-cifas
15. The environmental impact of AI is not just in carbon emissions, but also in their enormous water consumption: https://www.theregister.com/2026/03/10/us_datacenters_water_consumption/
16. AI shopping agents cannot be trusted to act in the best interests of customers: https://www.theregister.com/2026/03/10/cma_ai_agents/
17. Authors publish an empty book to protest the wholesale theft of their work to train AI: https://www.theguardian.com/technology/2026/mar/10/thousands-authors-publish-empty-book-protest-ai-work-copyright
18. AI needs to be backed-up by humans if you want to make quality decisions: https://www.stuff.co.nz/nz-news/360949122/man-banned-instagram-hours-after-posting-shirtless-image-himself-and-his-dog
19. A new AI trained on large datasets of DNA sequences could help design new life forms: https://www.hpcwire.com/bigdatawire/2026/03/05/a-new-ai-model-could-help-scientists-design-new-forms-of-life/
20. If an AI rewrites code, is it really new code? https: https://arstechnica.com/ai/2026/03/ai-can-rewrite-open-source-code-but-can-it-rewrite-the-license-too/
21. Top ten Python libraries for working with Large Language Model AI: https://www.kdnuggets.com/10-python-libraries-every-llm-engineer-should-know
22. Using AI for medical support on the International Space Station: https://www.extremetech.com/aerospace/nasa-trials-ai-medical-imaging-aboard-iss-to-reduce-reliance-on-ground
23. Microsoft wants to hire out its AI just like human workers: https://www.computerworld.com/article/4141904/microsoft-wants-you-to-hire-its-ai-agents.html
24. AI can now cooperate to break into systems and exfiltrate data: https://www.theregister.com/2026/03/12/rogue_ai_agents_worked_together/
25. Ai is antithetical to a education, especially in the humanities, as it cripples thinking in students: https://www.theguardian.com/technology/ng-interactive/2026/mar/10/ai-impact-professors-students-learning
26. Amazon now requires human sign-off of any changes made by an AI: https://arstechnica.com/ai/2026/03/after-outages-amazon-to-make-senior-engineers-sign-off-on-ai-assisted-changes/
27. Will lawyers be less threatened by AI than other professions? https: https://lawnews.nz/legal-profession/ai-is-a-language-based-revolution-lawfest-conference-told/